Adding a brief, hand-written video context improved LLM toxicity detection in Serbian, Croatian, and Bosnian, raising F1 scores by up to 0.10 across four models.
Facebook, Twitter struggling in fight against Balkan content violations,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Large Language Models for Toxic Language Detection in Low-Resource Balkan Languages
Adding a brief, hand-written video context improved LLM toxicity detection in Serbian, Croatian, and Bosnian, raising F1 scores by up to 0.10 across four models.